Glioma Classification of MR Brain Tumor Employing Machine Learning
Jyotsna Dogra1, Shruti Jain2, Meenakshi Sood3

1Jyotsna Dogra, Department of Electronics and Communication, Jaypee University of Information Technology, Solan, India.
2Shruti Jain, Department of Electronics and Communication, Jaypee University of Information Technology, Solan, India.
3Meenakshi Sood, Department of Electronics and Communication, Jaypee University of Information Technology, Solan, India.

Manuscript received on 02 June 2019 | Revised Manuscript received on 10 June 2019 | Manuscript published on 30 June 2019 | PP: 2676-2682 | Volume-8 Issue-8, June 2019 | Retrieval Number: H7543068819/19©BEIESP
Open Access | Ethics and Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)

Abstract: The remarkable performance achieved by machine learning for glioma classification has gained immense attention in the medical domain. The accurate knowledge of the glioma grading provides better treatment planning and diagnosis. In this research work a hybrid approach is proposed that integrates the Glioma segmentation and binary classification of the High- and Low-Grade Glioma. The proposed framework consists of several steps including targeted tumor segmentation, feature extraction, feature selection and classification using machine learning techniques (Support Vector Machine (SVM) and k-Nearest Neighbor (kNN)). An accurate segmentation of the targeted tumor region is obtained by applying the fuzzy clustering technique and the first order and second order statistical features are extracted from the complete imaging feature set. The most prominent features are selected using the t-test that are provided for performing the classification using SVM and kNN classifiers. The proposed hybrid framework was applied on a population of 300 MR brain tumor images diagnosed as 200 HGG tumors and 100 LGG tumors. The binary SVM and kNN classification, accuracy and performance metric is assessed by 10-fold cross validation. An accuracy of 94.9% and 91% is obtained for SVM and kNN classifiers respectively.
Keyword: Machine learning, classification, glioma, Magnetic Resonance Imaging (MRI), Accuracy.
Scope of the Article: DMachine Learning.